The comments from Collin and Rojo have some issues that seems to come up a lot in the analysis of the effect of the shoes (and their relation to other performance augmentations). The first few comments I have are in relation to the wind calculation and economy extrapolation, and while some of the issues come up with the shoe analysis, if you don't want to read that, just skip to the bottom about the shoes.
This is not correct. We actually have fewer studies on the effect of drag and horizontal force manipulation in running than we do the shoes (which woefully limited itself).
There are a lot of assumptions that went into that final number. Yes, wind speed profiles have a logarithmic profile that are valid over large distances through the atmosphere, but the effects of near-surface structures and the environment throw those off quite a bit, not to say anything about any effects of surrounding runners and the pack. Here's a good paper for wind conversions:
https://www.acoustics.asn.au/conference_proceedings/AAS2004/ACOUSTIC/PDF/AUTHOR/AC040116.PDFSimilarly, jumping from a windspeed estimation (which itself might have enormous variation) to a "second per mile" estimate is quite spurious. Not only would the effect itself be highly individualized, but the actual magnitude of the average effect is not well characterized.
Also note that horizontal forces applied during running are not linear - the benefit of a greater and greater tailwind taper off and approach an asymptote, whereas the cost of a greater and greater headwind exponentially increase.
A good paper on that:
https://journals.physiology.org/doi/full/10.1152/jappl.1999.86.5.1657All of that is simply to demonstrate that our simplifying assumptions to try and attach a number to some ecologically complex benefit - a tailwind at Boston - is really tricky, and might not be very informative when we try to compare it to other means of performance benefits.
I'm not sure that's true either - to say that the benefit one gets in his or her running economy does not change (from any source - wind, shoes, etc.) over the course of a marathon is not necessarily a safe assumption. We've seen evidence that running economy decreases during a marathon, so one could certainly posit that something that improves running economy may have a greater offset from the decrement if it's working to preserve some of the things that decay, like muscle function or strength (though that is also likely dependent on the mechanism - maybe wind is independent, but the shoes have a nonstationary benefit over 26.2 miles, or visa versa)
Also, even if it is distance-independent, it's not going to be speed-independent. Benefits to running economy decay at faster and faster speeds, so the tailwind to the person running 3:30 (3.3 m/s) is a substantial difference (much higher) to the person running 2:03 (5.7 m/s).
This is important if we are informing (or adjusting/checking) our performance-improvement estimates off of correlative data. What we observe in a bulk of runners around 3 hours, or even in runners we know around 2:30 or 2:20, does not necessarily extrapolate to the elites.
All of that being said, maybe 4:22 is a good estimate - maybe the wind was actually higher on the ground, but the effect and benefit of it is lower than what was assumed - but again, if it's gut-checked against a correlative measure (what we actually think Mutai was capable of or what we know other slower runners did run), it carries some fallacy that might be dangerous to compare. Maybe 2:10 is a better estimate though, or 1:00 or 3:15, or 5:20. It's probably not helpful to move the conversation forward to try and nail down a specific number there on something as complex as wind during a 26 mile race, and that applies to a lot of mechanisms.
All of that wasn't to knock down your analysis, but rather to highlight the trouble we get ourselves into if we try and attach concrete values to a multi-step extrapolation process. It can tie us to the presumed validity of the result, when there's error and variance in each step of the estimation that multiplies.
This is an issue that comes up again and again in the shoe debate - what previous runners in the recent past ran - and gets back to the issue and dangers of trying to compare performance effects.
When assessing the shoes against historical precedent at an elite level, there is an enormous confounding variable: change in PED policing
At the same time the shoes became widely available (2018), the first WADA-accredited blood lab was opened in East Africa (Kenya) and Kenya started massively ramping up its antidumping efforts under pressure from WA and WADA. So while we may have seen an enormous gain from the shoes at the top level, we also very likely saw a drop due to more stringent testing across the board. Just like there was a slight stagnation in the early 2010s, after the initial introduction of Biological Passport, this got many more athletes that routinely populate the top 20 and top 50 lists onto the passport and served to be even more stringent on the guys already at the top (the ones running 2:03 and 2:04 before the shoes). I'm certainly not going to say top guys are all totally clean now, but I suspect any benefits they're getting from any cheating is quite a bit smaller than it was 10 or 20 years ago.
So if we see a bunch of top-level guys winning races a 1:00-1:30 faster than they had been in 2012-2016, it's not a valid assumption that the benefit to an elite level 2:05-2:06 marathoner is 1:00-1:30. It might be 3:00-3:30, but they haven't been pharmaceutically augmented by 2:00. Again, that gets into some tricky business of mixing benefit estimations, but the point is that the trends we see at the top today are likely a combination of shoes (faster) and better PED policing (slower).
With respect to Kipchoge and Kamworor's WR in relation to the many of the names on that list, I think that might actually be very illustrative of that point. That list you provided has many names that highly suspect in my book. I don't think it's a stretch to say Kipchoge's best performance was his 2:04:00 in 2015, and that if he was still capable of that in Nike Streaks, or even declined slightly, say to 2:04:30, that the befit of the Next% is 2:30-3:00 to him. I think that's entirely logical.
I think one of the strange benefits of the shoes on both the roads and the track is that they're allowing us to re-write the record books in an age when we are at least a little bit more confident in PED policing measures. It's not perfect, but it's likely much less flagrant or of a smaller magnitude than in the past. It's a shame for a lot of national records and marks that were presumed clean (that were much slower than world records, haha), but I guess there are pros and cons for everything.
To put it all together, we have to be careful when our extrapolations are based on historical comparisons and correlations. Whether it's with runners at different speeds and abilities, or in runners of a different era.
Now, all of the being said - if the original intent was rooted in assessing Rupp's performance relative to Hall, I'd have to say we have pretty compelling evidence that Hall > Rupp at the marathon and half-marathon. Rupp is the best in the US right now, but the US is absolutely not world-class right now, so that means very little. Hall performed as well or better than Rupp over the HM and M distances in normal racing flats.